FITNESS FOR USE (Decision Rules and Target Measurement Uncertainty), Part-1 for Analytical life cycle management.

1. FITNESS FOR USE
Decision Rules and Target Measurement Uncertainty
Part-1
Chandra Prakash Singh

2. Reportable result is fit for use…………?
It is important, therefore, to understand what the result will be used for and to have a way of defining criteria
that can be used to assess the fitness of results for their purpose or use.
Product Analytical Test
Reportable
results
Reportable results are generated in
order to make decisions.
Decision
Rules
Driver for adopting the lifecycle approach to an analytical procedure is to ensure that the reportable result is fit for use.
The analytical target profile (ATP) concisely defines the requirements for a reportable result to be fit for use.

3. Target Measurement Uncertainty (TMU)
TMU which is defined in the International Vocabulary of Metrology as a “measurement uncertainty (MU) specified as an
upper limit and decided on the basis of intended use of measurement results.”
The TMU can become part of the ATP, which is valuable to the pharmaceutical industry because it provides a mathematical
proof that reportable results are suitable for use.
Decision Rule (DR)
The development of decision rule (DR) concepts provides an approach that can be helpful for determining fit for use.
 Decision rules, which have been used to provide organizations with procedures for accepting and rejecting products.
 The decision rule (DR) defines the use of the reportable result, and can provide the information, such as acceptable
probabilities, needed to set the target measurement uncertainty (TMU).
The Relation to the Analytical Procedure
DR can provide an understanding of the maximum variability or maximum uncertainty (or TMU) that can be associated
with a reportable result, and whether that reportable result may still be fit for its intended use.

4. Fit for Use and the Measurand
 The DR approach allows the defining of any decision that will be made with the reportable result.
 The process for defining the DR must include the end user of the reportable result.
End users may include the lead in a clinical study, the production manager, the stability study coordinator, the regulator
releasing a lot of drug product (DP), or other positions.
The measurand helps to determine the intended use for the procedure and should include important aspects and
conditions, such as the following:
 Description of the quantity to be measured, more commonly known as the units;
 The analyte, which is the entity actually being measured;
 The matrix, if relevant;
 Whether the reportable result refers to the laboratory sample or to the “parent body”;
 If relevant, the measurand may include information on the analytical procedure itself. For operationally defined
(empirical) procedures, this is often relevant. For example, experimental conditions, such as the temperature for loss on
drying, may be specified.

5. Another challenge in defining fit for use is that there may be more than one use for a reportable result. Thus, when
developing criteria for the reportable result, we need to consider all possible uses, including release, stability, and others,
as each has its own need for the data.
The required quality or tolerance for MU (bias and uncertainty) associated with the reportable result may be different for
each use. The uncertainty that is acceptable for the reportable result to release a lot may be larger than that required for
the stability study of that lot.
Following a defined process using knowledge management, risk analysis, and process mapping helps define the uses of
the reportable result and its required quality. Each use will require its own DR and ATP because the measurand and
acceptable risks may be different.

6. Any discussion on introducing the concept of DRs in the pharmaceutical industry inevitably must also consider the process by
which specifications are set.
In present practice, pharmaceutical specifications are established to ensure safety and efficacy and are in fact generally tighter
than the specifications required for safety and efficacy. This is because of quality rationales related to the capability of
processes and analytical procedures.
These rationales may not involve the actual performance of the DP in the clinical environment, or this connection is not
always apparent. With quality by design (QbD), there is a desire to more closely link the acceptance criteria in specifications
to the clinical effect.
By ensuring that specifications are based on clinical relevance. DRs can act as the link between the reportable result and the
clinical effect of the DP.
The Link of Decision Rules with Specifications

7. The Link of Decision Rules with Specifications
The requirements for the clinical use of a Drug Product (DP) may be much wider than the specifications of the DP. This
relation between clinical requirements and product specifications may not be apparent. A DR based on the intended use
of the reportable result can be used to make this relationship apparent.
Clinical Requirements
Specification Zone for
Drug Product
Upper Specification Limit
Lower Specification Limit

8. A DR is a documented rule that describes how the MU will be allocated with regard to accepting or rejecting a product
according to its specification and the result of a measurement.
DRs are developed using risk - assessed by considering the potential harms and hazards - and probability.
DRs give a prescription for the acceptance or rejection of a product based on the measurement result, its uncertainty, and
the specification limit or limits.
Additionally, DRs can take into account an acceptable level with regard to the probability of making a wrong decision.
The wrong decision can lead to accepting an out-of specification (OOS) reportable result which is not true or rejecting an
OOS reportable result which is true, or a false failure or a missed fault.
DECISION RULES

9. The Decision Rule and Compliance
The four possible outcomes when a reportable result and its expanded uncertainty are compared with a specification. Each
outcome shows a normal distribution curve with the reportable result at the center and the width of the distribution
determined by the uncertainty.
Upper Limit
1. Result is above
the limit. Limit is
below expanded
uncertainty.
2. Result is above
the limit. Limit is
within expanded
uncertainty.
3. Result is below
the limit. Limit is
within expanded
uncertainty.
4. Result is below
the limit. Limit is
below expanded
uncertainty.

10. A normal distribution curve centered on the reportable result shown by the X. For this reportable result, using a coverage factor
of 1.65 with a 95% level of confidence, the probability of making a wrong decision is 5%. This is for a reportable result that is
1.65 × standard uncertainty below the upper specification limit.
Expanded Uncertainty and Coverage Factors
Concentration Limit
Using a coverage factor, k=1.65, with
a 95% confidence, the probability of
making a wrong decision is 5%.
The reportable result, indicate by
the X, is 1.65µ, below the upper
specification limit.
X
A coverage factor is similar to the z factor for a standard normal distribution and is typically in the range of two to three (e.g.,
for a 95% level of confidence, a k = 2 is used).

11. Types of Decision Rules (DRs)
Decision Rules (DRs)
 Acceptance zones to define the range within which the product will be accepted.
 Rejection zones to define the range within which it will be rejected.
 Transition zone defines a range within which the product is not immediately accepted or rejected; instead, other activities
may be performed, such as additional testing, testing using a different technique, or an investigation. These zones may or
may not align with the specification ranges.
Different types of DRs use different acceptance, rejection, or transition zones to fully define the subsequent decision.
The end user of the data can judge the impact of being wrong 5% of the time. If the impact is not acceptable, a different level of
probability can be chosen, the uncertainty can be reduced, or the specification limits can be changed.

12. Specification Zone
Upper Specification Limit
Lower Specification Limit
Acceptance Zone Rejection Zone
Rejection Zone
Simple DRs: Simple Acceptance or Rejection Zone
Product conformance is verified if the measurement result is in the specification zone; otherwise rejection is verified.
The specification and acceptance zones are aligned, so the acceptance zone is called the “Simple Acceptance Zone” and the
range outside the specification zone is called the “Simple Rejection Zone”.
Based on the risk, the acceptance and rejection zones may be different from those of the specification.
These DRs use guard bands, which is the magnitude of the offset from the specification limit to the acceptance or rejection
zone boundary.

13. A DR that uses a stringent acceptance zone and relaxed rejection zones. The guard bands, gL and gU, are within the
specification zone. For this DR, a product will be accepted if the risk of being OOS is low.
gL
Specification Zone
Upper Specification Limit
Lower Specification Limit
Stringent
Acceptance
Zone
Relaxed
Rejection
Zone
Relaxed
Rejection
Zone
gU
Relaxed Rejection Zone
The rejection zone is increased and is inside the specification zone by the amount of a guard band (gL and gU).
Stringent Acceptance Zone
A specification zone that is reduced from the upper and lower specification limits by a guard band
(often called an “alert” or “internal release limit” in the pharmaceutical industry).

14. A DR that uses a relaxed acceptance zone and stringent rejection zones. The guard bands, gL and gU, are outside the
specification zone. For this DR, a product will be rejected only if the risk of the product being OOS is high.
gL
Specification Zone
Upper Specification Limit
Lower Specification Limit
Relaxed
Acceptance
Zone
Stringent
Rejection
Zone
gU
Stringent
Rejection
Zone
Relaxed Acceptance Zone
The acceptance zone is increased beyond the specification zone by a guard band.

15. A DR that uses a transition zone. If a reportable result is obtained within the transition zone, additional activities prescribed in
the DR can be undertaken. For example, an investigation may be initiated to determine the reason for obtaining a value in the
transition zone and additional tests may be prescribed, or a different procedure with a lower uncertainty may be used.
gL
Specification Zone
Upper Specification Limit
Lower Specification Limit
Stringent
Acceptance
Zone
Rejection
Zone
Rejection
Zone
gU
Transition
Zone
Transition
Zone
Stringent Rejection Zone
The rejection zone is increased beyond the specification zone by a guard band.
Transition Zone
Area between the acceptance zone and rejection zone.

16. The MU should meet the requirements of the acceptance zone. The TMU is set using the acceptance zone and the
acceptable probability of obtaining reportable results within the transition zone or rejection zone. Setting the TMU in this
way ensures that the reportable result will be fit for purpose, and allows the TMU to become part of the ATP.
Present Practice in the Pharmaceutical Industry
 Current practice in the pharmaceutical industry may be described as using two types of DRs: “simple” DRs and DRs that
use transition zones.
 Applying DR theory to present practice allows for a better understanding of risk and probability and can also define
additional criteria, such as the TMU, for the ATP.
 A majority of USP monograph specifications are simple DRs. When a reportable result is obtained that is at the
specification limit, there is a 50:50 probability that the true value is OOS. If this probability may not be acceptable, a
transition DR that uses a guard band and transition zone can be used. In current practice these are called internal release or
alert limits.

17. A USP specification of 98.0%–102.0%. If a reportable result is at or close to 98.0%, there is a 50:50 probability
that the true value is below the lower limit.
96
95 97 98 99 100 104 105
103
101 102
Lower Limit Upper Limit
50% Probability the true value
is above the lower limit.
50% Probability the true value
is below the lower limit.

18. ?
Thanks